Proc. Natl. Acad. Sci. USAVol. 90, pp. 9552-9556, October 1993Cell Biology

Mechanism of mitotic block and inhibition of cell proliferation bytaxol at low concentrations

(HeLa cells/microtubule dynamics/vinblastine)

MARY ANN JORDAN*, ROBERT J. Toso, DOUG THROWER, AND LESLIE WILSONDepartment of Biological Sciences, University of California, Santa Barbara, CA 93106-%10

Communicated by John Carbon, July 14, 1993 (received for review June 8, 1993)

ABSTRACT Taxol inhibited HeLa cell proliferation byinducing a sustained mitotic block at the metaphase/anaphaseboundary. Half-maximal inhibition of cell proliferation oc-curred at 8 nM taxol, and mitosis was half-maximally blockedat 8 nM taxol. Inhibition of mitosis was associated withformation of an incomplete metaphase plate of chromosomesand an altered arrangement of spindle microtubules thatstrongly resembled the abnormal organization that occurs withlow concentrations of vinblastine and other antimitotic com-pounds. No increase in microtubule polymer mass occurredbelow 10 nM taxol. The mass of microtubules increasedhalf-maximally at 80 nM taxol and attained maximal levels (5times normal) at 330 nM taxol. At submicromolar concentra-tions, taxol suppressed growing and shortening at the ends ofmicrotubules reassembled in vitro from bovine brain tubulin ina manner that resembled suppression by vinblastine. Taxol wasconcentrated in HeLa cells several hundredfold to levels thatwere similar to those which suppressed dynamic instability invitro. The results indicate that taxol shares a common antipro-liferative mechanism with vinblastine. At its lowest effectiveconcentrations, taxol appears to block mitosis by kineticallystabilizing spindle microtubules and not by changing the massof polymerized microtubules.

The antimitotic antitumor drug taxol has undergone exten-sive clinical development as a result of its efficacy in thetreatment of refractory ovarian cancer and its potential valuefor the treatment of breast, lung, and other cancers (1). Themechanism of action of taxol has been considered to beunique. The target for taxol appears to be microtubules, andin contrast to vinblastine, colchicine, and other antimitoticcompounds that can inhibit microtubule polymerization bothin vitro and in cells, taxol can enhance microtubule polymer-ization (2, 3). Taxol can block mitosis, induce extensiveformation of microtubule bundles in cells, and induce multi-nucleation of cells during interphase (4-6). Enhanced micro-tubule polymerization has been suggested to be responsiblefor the antitumor activity of the drug.However, in recent studies, we found (7-9) that at their

lowest effective concentrations, vinblastine and vincristineinhibit mitosis and cell proliferation in HeLa cells withoutdecreasing the mass of microtubules and with only subtlechanges in the organization of the mitotic spindles. We alsofound that low concentrations of vinblastine inhibit dynamicinstability and treadmilling of reassembled bovine brain mi-crotubules in vitro without appreciably affecting the micro-tubule polymer mass (10, 11). Dynamic instability consists oftransitions between phases of growing and phases of short-ening at microtubule ends. Treadmilling is the net addition oftubulin at one microtubule end and the balanced net loss oftubulin at the opposite microtubule end. Both kinds of

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

dynamics have been shown to occur in mitotic spindlemicrotubules, which are characterized by very rapid dynamicbehaviors (for review, see ref. 12). Mitotic block at lowconcentrations of vinblastine is thought to occur by suppres-sion of tubulin exchange at ends of mitotic spindle microtu-bules (7, 8, 11).

In the present study, we found that low concentrations oftaxol inhibited mitosis in HeLa cells and suppressed dynamicinstability at the ends of bovine brain microtubules in vitro inthe absence of enhanced microtubule polymerization. Inaddition, inhibition of mitosis was associated with an alteredorganization of mitotic spindles that was strikingly similar tothat induced by the vinca alkaloids and several other anti-mitotic drugs, suggesting that mitotic block and inhibition ofproliferation by all of these drugs, including taxol, at theirlowest effective concentrations, involves kinetic stabilizationof mitotic spindle microtubule dynamics rather than alter-ations in microtubule polymer mass.

MATERIALS AND METHODSHeLa S3 cells (American Type Culture Collection) weregrown in monolayers at 37°C without antibiotics in 5%C02/95% air (7). Cell proliferation was determined by count-ing cells by hemocytometer at the time of taxol addition and20 h later. Mitotic index, cell morphology, and spindleinterpolar distances were determined by immunofluores-cence microscopy (8). Levels of polymerized tubulin in cellswere determined by measuring the tubulin content of isolatedstabilized cytoskeletons by an ELISA (two to six determi-nations per taxol concentration) (13).For analysis of the effects of taxol on dynamic instability,

microtubules were assembled to steady state by addingbovine brain tubulin (depleted of microtubule-associatedproteins, >99% tubulin, 1.5 mg/ml in 100 mM Mes/1 mMEGTA/1 mM MgSO4/1 mM GTP, pH 6.7, 30°C) in thecontinuous presence of taxol onto the ends of Strongylocen-trotus purpuratus sperm flagella axonemal seeds (11). Underthe conditions used, control microtubules grow mainly at theplus ends of the seeds. Microtubules at the minus ends werefew in number and very short. In the presence of taxol at theconcentrations used, the microtubules at the two ends of theseeds were indistinguishable in length, number, and dynam-ics, and thus the measurements represent a mixture ofmicrotubules at both ends of the seeds. Measurements ofmicrotubule lengths were made at 15-s intervals from imagescaptured using a Hamamatsu C2400 (Newvicon) video cam-era with video-enhanced differential interference contrastmicroscopy on a Zeiss IM35 microscope with a temperature-controlled stage (11). At least 100 measurements of anaverage of seven microtubules from a minimum of twoexperiments were used for each experimental condition.

*To whom reprint requests should be addressed.

9552

Proc. Natl. Acad. Sci. USA 90 (1993) 9553

Taxol uptake into HeLa cells was determined after incu-bating cells for 20 h with [3H]taxol (3-1000 nM; specificactivity, 4-3600 Ci/mol; 1 Ci = 37 GBq) (7). Taxol and[3H]taxol were gifts from the National Cancer Institute([3'-3H]taxol, NSC 125973) or were purchased (MoravekBiochemicals, La Brea, CA). [14C]Hydroxymethylinulin (5.4,uM; specific activity, 10 Ci/mol; New England Nuclear) wasadded to cell suspensions before collection to determineextracellular volume.

RESULTSRelationship of Inhibition of Cell Proliferation, Mitotic Block,

and Enhancement of Microtubule Polymer Levels by Taxol.HeLa cells were incubated for the duration of one cell cyclewith taxol over a broad range of concentrations. After 20 h,cytoskeletons were isolated to determine the mass of tubulinin the form of microtubules. In parallel experiments, inhibitionof proliferation and mitotic indices were determined. Taxolinhibited cell proliferation half-maximally at a concentration of8 nM, and inhibition was complete at concentrations >33 nM(Fig. 1A). Taxol induced the accumulation of cells in mitoticmetaphase half-maximally at a concentration of -8 nM, andmaximal mitotic accumulation (80-95%) occurred at taxolconcentrations of 33 nM and above. Thus mitotic accumula-tion occurred in parallel with inhibition of proliferation. Noincrease in microtubule polymer mass occurred at <10 nMtaxol. The mass of microtubules then increased as the taxolconcentration was raised, attaining a half-maximal increase ata concentration of 80 nM and a maximal increase of 500% ofthe normal level at 330 nM taxol.

14

ca)0

ci)

1U

a0

L-

600

10[Taxol], nM

0

cnU)

co

2c)

0

0

FIG. 1. (A) Taxol concentration dependence for metaphase arrest(open circles, left axis; arrow denotes control value), for inhibitionof proliferation (solid circles, left axis), and for the increase inmicrotubule mass (open squares, right axis). (B) Taxol concentrationdependence of spindle organization in HeLa cells after incubationwith taxol for one cell cycle. (A) Accumulation of cells in metaphasewas concomitant with inhibition of proliferation but was accompa-nied by little or no increase in the mass of microtubule polymer. (B)Percentages of metaphases that were normal (open circles), abnor-mal bipolar types I or II (solid squares), and ball-shaped type III(open diamonds). Arrows denote control values for normalmetaphases (upper arrow), abnormal bipolar (middle arrow), and ball(lower arrow).

Cells accumulated in metaphase or in a metaphase-likeconfiguration (see below), but no cells were in anaphaseeither after a 20-h incubation with 10 nM taxol (Table 1) orduring long-term incubation (as long as 48 h) with 10 nM taxol(data not shown). Thus accumulation in mitosis represents asustained block at the metaphase/anaphase boundary. Of thecells that were in interphase at low taxol concentrations (3-10nM), a large percentage (31-38%) consisted of cells with two,three, or more nuclei (Table 1). These results suggest that themitotic block induced by low taxol concentrations is not assustained as at higher taxol concentrations; some cells canescape from mitotic block and become abnormal multinucle-ate cells.

Effects of Taxol on Spindle Organization. The arrangementsof microtubules, chromosomes, and centrosomes of controlcells and of cells incubated with taxol are shown in Fig. 2.Mitotic spindles of cells blocked in metaphase by low con-centrations of taxol strongly resembled spindles of cellsblocked by low concentrations of other antimitotic drugsincluding vinblastine, vincristine, nocodazole, colchicine,and podophyllotoxin (7, 8). Some spindles were blocked in abipolar configuration that was normal (data not shown);chromosomes were all contained in a compact metaphaseplate. Other spindles (20-32% with 0.33-10 nM taxol) wereblocked in a nearly normal configuration; these spindles werebipolar with a compact metaphase plate of chromosomes butwith some chromosomes located near the spindle poles (Fig.2 d-J) (spindle types I and II, ref. 8). Astral microtubules intaxol-blocked spindles appeared more prominent than incontrol spindles, and the central spindle (the distance frompole to pole) was shorter than in control spindles. Forexample, the length of the central spindle was reduced from7.4 ± 0.2 ,m in control spindles to 4.0 ± 0.4 um in bipolarspindles by 10 nM taxol (Table 1).With increasing concentrations of taxol, spindle morphol-

ogy became more abnormal; increasing numbers of chromo-somes were located near the poles of bipolar spindles ratherthan in the metaphase plate (data not shown). Also as thetaxol concentration was increased, many spindles had nobipolar organization but were ball-shaped aggregations ofcondensed chromosomes containing one or more asters ofmicrotubules (Fig. 2 g-i). These resembled classical colchi-cine mitoses (C-mitoses of ref. 14) or type III aberrantspindles (8). The decrease in the proportions of bipolar(normal and aberrant types I and II) spindles and the increasein proportion of ball-shaped (type III) spindles with increas-ing taxol concentration are shown in Fig. 1B.

Table 1. Effects of taxol (20-h incubation) onmetaphase/anaphase transition, multinucleation, and spindlepole separation

Cells in anaphase/ Multinucleated InterpolarTaxol, cells in metaphase, interphase distance,nM no./no. cells, % Am0 0.14 ± 0.03 2.7 ± 0.7 7.4 + 0.20.3 0.09 ± 0.04 3.2 ± 1.2 6.6 ± 0.11 0.11 ± 0.05 11.6 ± 4.0 6.9 ± 0.13 0.04 ± 0.04 31.2 ± 17.9 6.1 ± 0.56 0.007 ± 0.004 32.6 ± 12.610 0 37.7 ± 11.5 4.0 ± 0.433 0 16.5 4.5 2.8 ± 0.2100 0 14.0 6.81000 0 7.3 2.3

Data are the mean ± SEM ofthree experiments for the ratio of cellsin anaphase to metaphase and the percent multinucleated interphasecells (multinucleated cells have more than one nucleus). The inter-polar distance is the distance between centrosomes at oppositespindle poles (mean + SEM of .25 spindles).

Cell Biology: Jordan et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

FIG. 2. Microtubules (a, d,and g), chromosomes (b, e, andh), and centrosomes (c,f, and i) ofHeLa cell mitotic spindles afterincubation for 18-20 h with taxol.(a-c) Control cell spindle with fewastral microtubules and a well-defined compact metaphase plateof chromosomes. (d-f) At 6 nMtaxol, an abnormal bipolar spindle(type I) with prominent astral mi-crotubules (arrow in d) and chro-mosomes near the spindle poles(arrows in e). (g-i) At 1 ,uM taxol,a ball-shaped chromosomal masswith a monopolar microtubule andcentrosome arrangement (typeIII). (Bar = 10 um.)

The morphological changes in spindle structure induced bytaxol were nearly identical to those that occurred with otherantimitotic drugs (7, 8). However, there were some minordifferences. Some bipolar spindles blocked by low concen-trations of taxol appeared to have reduced numbers ofinterpolar microtubules (data not shown), in contrast with thebipolar spindles induced by low concentrations of otherantimitotic drugs that appeared to have normal numbers ofinterpolar microtubules. This observation with taxol is con-sistent with the finding that addition of 10-20 AM taxol toPtKl cells in early anaphase caused the disappearance ofmost interzonal microtubules within 5 min (15). In addition,in the present work upon incubation with 10-100 nM taxol,some mitotic asters contained no centrosomes (data not

shown). Similar results were obtained in PtK2 cells afterincubation with micromolar concentrations of taxol (16, 17).

Induction of Microtubule Bundling. Microtubule bundlesdid not form in the taxol concentration range in which littleor no increase in microtubule mass occurred (1-10 nM).However, with 10 nM taxol, microtubules often becameoriented in parallel fashion. (Compare the meshwork ofmicrotubules of a control cell in interphase in Fig. 3a with thearray of parallel microtubules radiating out from the nucleusafter incubation with 10 nM taxol in Fig. 3b.) Loosely packedbundles of microtubules were observed in a few cells incu-bated with 33 nM taxol (Fig. 3c), a concentration at which themicrotubule polymer mass was double that of control cells.Massive bundles of microtubules formed at higher taxolconcentrations (e.g., 1 ,M, Fig. 3d).

FIG. 3. Microtubules of HeLa cells in interphase incubated for 18-20 h with taxol. (a) Control irregular meshwork of microtubules. (b) At10 nM taxol, a somewhat parallel alignment of microtubules but absence of microtubule bundles. (c) At 33 nM taxol, a loosely packed bundleof microtubules. (d) At 1 ,uM taxol, three compact bundles of microtubules. (Bar = 10 Am.)

9554 Cell Biology: Jordan et al.

Proc. Natl. Acad. Sci. USA 90 (1993) 9555

Table 2. Suppression of microtubule dynamic instability by taxolGrowing rate, Shortening rate, Transition Dynamicity,

Taxol, tubulin tubulin % time in frequency from tubulin,uM dimers per s dimers per s attenuation G or S to A dimers per s

0 (control) 106 ± 13 104 ± 16 19 ± 3 0.65 83 ± 80.1 59±6 55±4 45±1 1.6 31±40.5 48 ± 3 49 ± 5 53 ± 8 2.3 23 ± 2

Attenuation is a phase of no detectable length change at the microtubule ends. Transition frequencyis the number of transitions divided by the total time the microtubules spent in the phases from whichthe transition occurred (11). G, growing; S, shortening; A, attenuation. Dynamicity is the total tubulinexchange in tubulin dimers per s (11). Control data are the same as in ref. 11.

Suppression of Dynamic Instability at Microtubule Ends inVitro. Video microscopy was used to determine the effects oftaxol on dynamic instability at ends ofindividual bovine brainmicrotubules. Taxol (0.1 and 0.5 tkM) significantly inhibitedthe microtubule shortening rate and the microtubule growingrate and strongly inhibited the overall exchange of tubulin atmicrotubule ends (dynamicity, Table 2). Taxol increased thepercentage of total time that the microtubules remained in anattenuated state, neither growing nor shortening detectably,from 19% in controls to >50% at 0.5 ALM taxol. Taxol alsoincreased the frequency of transition or switching fromgrowing or shortening to attenuation. The suppression ofdynamic instability by 0.1 and 0.5 ,uM taxol occurred withoutany detectable increase in the mass of polymerized tubulin(an increase in sedimentable microtubule polymer was de-tected under the conditions of the video microscopy only attaxol concentrations of 21 uM; W. B. Derry, L.W., andM.A.J., unpublished data).

Intracellular Taxol Concentration. Fifty percent mitoticblock occurred at an added taxol concentration of 8 nM, aconcentration that is far below that necessary to suppressmicrotubule dynamics in vitro. Thus, we measured the in-tracellular taxol concentration after incubation of cells withradiolabeled taxol at concentrations that induced mitoticblock (Table 3). At all concentrations, taxol was concentratedseveral hundredfold intracellularly, to micromolar levels. Forexample, with 10 nM taxol, which induced 67 ± 24% mitoticblock, 96% aberrant spindles, and multinucleation of 38% ofthe cells that were in interphase (Fig. 1A and Table 1), thedrug was concentrated 480-fold to an intracellular level of 4.8± 0.7 AM. Therefore, while the subcellular location of all thetaxol is not completely known (18), the overall intracellulartaxol concentration was sufficient to suppress microtubuledynamics.

DISCUSSIONVery low concentrations of taxol are sufficient to inhibitproliferation of HeLa cells. Both half-maximal inhibition ofproliferation and 50% blockage in mitotic metaphase oc-curred at 8 nM taxol. The degree ofmetaphase block by taxolparalleled the degree of inhibition of cell proliferation at alltaxol concentrations. Inhibition was associated with forma-tion of an incomplete metaphase plate of chromosomes andan arrangement of spindle microtubules that strongly resem-bled the abnormal organization that occurs with low concen-

Table 3. Intracellular taxol concentration in HeLa cellsTaxol in medium Taxol in cells Fold Tubulin inat time 0, ,uM at 20 h, ,uM uptake polymer, /AM

0.003 1.8 ± 0.3 600 6.3 ± 0.80.01 4.8 ± 0.7 480 8.4 ± 1.50.1 40.5 ± 8.0 405 19.1 ± 1.81.0 111.0 ± 13.8 111 28.8 ± 2.9

Amount of tubulin in polymer is from data of Fig. 1A, assumingtotal tubulin at 2 mg/ml and 31-36% in polymer in control cells (7).

trations of vinblastine and other antimitotic drugs (8). Themost sensitive inhibitory effects oftaxol on proliferation werenot associated with an increase in microtubule polymer massor with the formation of microtubule bundles, actions oftaxolthat occur at relatively high drug concentrations. No increasein microtubule polymer mass occurred below a taxol con-centration of 10 nM (Fig. 1A), and 80 nM taxol was requiredto induce a half-maximal increase in the microtubule polymermass. Thus these results indicate that the most sensitiveaction of taxol on HeLa cell proliferation involves blockageof cell cycle progression at the metaphase/anaphase transi-tion in the presence ofa normal mass ofmicrotubule polymer.Submicromolar concentrations of taxol significantly al-

tered dynamic instability of microtubules in vitro. At plusends, taxol inhibited both the rates of growing and shorten-ing, strongly increased the percentage of time the microtu-bules remained in an attenuated state (neither growing norshortening detectably), and increased the frequency of tran-sition into periods of no detectable length change. Overall,taxol (0.5 uM) inhibited dynamicity by 70% (Table 2).

Interestingly, the inhibitory effects of taxol on dynamicinstability qualitatively resemble those of vinblastine at thekinetically rapid microtubule plus ends (11). Like taxol,vinblastine strongly suppresses the rates of growing andshortening, increases the percentage of time that the micro-tubules spend in an attenuated state, increases the frequencyof transition from growing or shortening to attenuation, andstrongly suppresses dynamicity. Thus, while vinblastine andtaxol interact with microtubules by different molecular mech-anisms (3, 19-21), both drugs kinetically stabilize microtu-bule ends in a remarkably similar fashion. Thus, the dataindicate that at low drug concentrations, taxol may share acommon antiproliferative mechanism with vinblastine inHeLa cells, namely, inhibition of mitosis by suppressing thedynamics of mitotic spindle microtubules.There are several mechanisms by which aberrant spindle

morphology and mitotic block could result from stabilizationof microtubule dynamics by taxol. For example, duringprometaphase, the plus ends of dynamic spindle microtu-bules appear to probe the cytoplasm until linkage with achromosomal kinetochore is established (22, 23). Reducedmicrotubule dynamics in the presence of taxol could result inthe attachment of fewer than a normal number of microtu-bules to the kinetochores, as occurs with vinblastine (9),which may lead to failure of chromosome congression (Fig.2 d-i). In addition, we were surprised to find, given the abilityof taxol at high concentrations to enhance microtubule po-lymerization, that taxol, like vinblastine, nocodazole,podophyllotoxin (8), and estramustine (24), induced short-ening ofthe microtubules ofthe central spindle and decreasedseparation of spindle poles (Table 1). Microtubules of thecentral spindle region flux or treadmill in a poleward directionduring metaphase and anaphase (25, 26). Taxol, like vinblas-tine, colchicine, podophyllotoxin, and nocodazole, inhibitstreadmilling or flux of microtubules in vitro (refs. 10 and27-29, and L.W., unpublished data), and shorter centralspindles may be due to a reduced flux rate. In addition, the

Cell Biology: Jordan et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

transition from metaphase to anaphase may require dynamicspindle microtubules for correct spindle formation, for chro-mosome separation, or for a thus far unknown signalingmechanism.At taxol concentrations >10 nM, the mass of microtubule

polymer in cells increased, attaining levels that were 500% ofcontrols at high taxol concentrations (330 nM). Microtubulebundles became prominent in this concentration range (Fig.3). Increased microtubule polymer mass and induction ofmassive bundles of microtubules may contribute to theantiproliferative action of taxol at high concentrations. How-ever, increased polymer mass and induction of microtubulebundling did not occur at the lowest effective concentrationsof taxol, so neither action of the drug can account forinhibition of proliferation by low concentrations of the drug.It is curious that even with a large excess of stabilizedmicrotubules, the interphase microtubule cytoskeleton candepolymerize or rearrange sufficiently for the cell to entermitosis and construct a mitotic-like spindle. In addition, itappears that either the dynamics of mitotic microtubules aremore sensitive than those of interphase microtubules tostabilization by taxol or that stabilization of microtubuledynamics in interphase does not prevent cell cycle progres-sion through interphase in HeLa cells.At high concentrations, taxol binds stoichiometrically to

tubulin in microtubules (3, 21). However, taxol suppressesdynamic instability of bovine brain microtubules in vitro attaxol/tubulin ratios as low as 1:150 (Table 2). Under condi-tions of the most sensitive mitotic block in the present work(3-10 nM taxol), the intracellular taxol concentration is40-70% less than the concentration of tubulin in microtubulepolymer (Table 1). Although the mechanism of inhibition ofmicrotubule dynamics by taxol has not been elucidated, theseresults indicate that inhibition of microtubule dynamics andinhibition of mitosis may result from the binding of smallnumbers of taxol molecules per microtubule.Taxol shares with vinblastine, nocodazole, colchicine, and

podophylloxin the ability to block mitosis at low drug con-centrations in the presence of normal amounts ofmicrotubulepolymer and with a similar aberrant spindle organization (8,12). All five compounds inhibit treadmilling of microtubulesin the absence of changes in the polymer mass (refs. 10 and27-29 and L.W., unpublished data). In addition, taxol, vin-blastine, nocodazole, and colchicine suppress microtubuledynamic instability in the absence of significant changes inthe microtubule polymer mass (Table 2 and refs. 11, 12, and30). Thus, the similar actions of these drugs on microtubulesin vivo and in vitro suggest that low concentrations of taxol,like low concentrations of other antimitotic drugs, blockmitosis and inhibit cell proliferation by inhibiting the dynam-ics of spindle microtubules.The taxol concentrations used in the present study are

considerably lower than those currently used clinically; e.g.,a steady-state plasma taxol concentration of 0.45 ,uM wasreported during a 24-h continuous intravenous infusion of 170mg oftaxol per m2 (31). These results, as well as evidence thatlow concentrations of taxol enhance the cytotoxicity ofestramustine (32), suggest that therapeutic administration oflower taxol concentrations than presently used might effec-tively inhibit tumor cell growth. In support ofthis idea, recentevidence indicates that mitotic block induced in HeLa cellsby taxol at low concentrations leads not only to inhibition of

cell proliferation but also to cell killing (M.A.J. and L.W.,unpublished data).

We thank Ms. Kim Wendell, Ms. Sara Gardiner, and Mr. HerbMiller for excellent technical assistance and Mr. Brent Derry forcareful reading of the manuscript and for measurements of criticalconcentration. This work was supported by Grant CA57291-01 fromthe National Cancer Institute (M.A.J.) and by Grant DHP-43E fromthe American Cancer Society (L.W.).

1. Rowinsky, E. K. & Donehower, R. C. (1991) Pharmacol. Ther.52, 35-84.

2. Schiff, P. B., Fant, J. & Horwitz, S. B. (1979) Nature (London)277, 665-667.

3. Parness, J. & Horwitz, S. (1981) J. Cell Biol. 91, 479-487.4. Fuchs, D. A. & Johnson, R. K. (1978) Cancer Treat. Rep. 62,

1219-1222.5. Schiff, P. B. & Horwitz, S. B. (1980) Proc. Natl. Acad. Sci.

USA 77, 1561-1565.6. Rowinsky, E. K., Donehower, R. C., Jones, R. J. & Tucker,

R. W. (1988) Cancer Res. 48, 4093-4100.7. Jordan, M. A., Thrower, D. & Wilson, L. (1991) Cancer Res.

51, 2212-2222.8. Jordan, M. A., Thrower, D. & Wilson, L. (1992) J. Cell Sci.

102, 401-416.9. Wendell, K. L., Wilson, L. & Jordan, M. A. (1993) J. Cell Sci.

104, 261-274.10. Jordan, M. A. & Wilson, L. (1990) Biochemistry 29, 2730-2739.11. Toso, R. J., Jordan, M. A., Farrell, K. W., Matsumoto, B. &

Wilson, L. (1993) Biochemistry 32, 1285-1293.12. Wilson, L. & Jordan, M. A. (1993) in Microtubules, ed. Hyams,

J. & Lloyd, C. (Wiley, New York), in press.13. Thrower, D., Jordan, M. A. & Wilson, L. (1991) J. Immunol.

Methods 136, 45-51.14. Levan, A. (1938) Hereditas 24, 471-486.15. Amin-Hanjani, S. & Wadsworth, P. (1991) Cell Motil. Cytoskel-

eton 20, 136-144.16. DeBrabander, M., Geuens, G., Nuydens, R., Willebrords, R. &

De Mey, J. (1981) Proc. Natl. Acad. Sci. USA 78, 5608-5612.17. Kallajoki, M., Weber, K. & Osborn, M. (1992) J. Cell Sci. 102,

91-102.18. Manfredi, J. J., Parness, J. & Horwitz, S. B. (1982) J. Cell Biol.

94, 688-696.19. Jordan, M. A., Margolis, R. L., Himes, R. H. & Wilson, L.

(1986) J. Mol. Biol. 187, 61-73.20. Singer, W. D., Jordan, M. A., Wilson, L. & Himes, R. H.

(1989) Mol. Pharmacol. 36, 366-370.21. Diaz, J. F. & Andreu, J. M. (1993) Biochemistry 32, 2747-2755.22. Rieder, C. L., Alexander, S. P. & Rupp, G. (1990) J. Cell Biol.

110, 81-95.23. Hayden, J. H., Bowser, S. S. & Rieder, C. L. (1990) J. Cell

Biol. 111, 1039-1046.24. Sheridan, R. V., Speicher, L. A. & Tew, K. D. (1991) Eur. J.

Cell Biol. 54, 268-276.25. Mitchison, T. (1989) J. Cell Biol. 45, 515-527.26. Mitchison, T. J. & Salmon, E. D. (1992) J. Cell Biol. 119,

569-582.27. Wilson, L., Miller, H. P., Farrell, K. W., Snyder, K. B.,

Thompson, W. C. & Purich, D. L. (1985) Biochemistry 24,5254-5262.

28. Wilson, L. & Farrell, K. W. (1986) Ann. N. Y. Acad. Sci. 466,690-708.

29. Jordan, M. A. & Farrell, K. W. (1983) Anal. Biochem. 130,41-53.

30. Wilson, L., Toso, R. J. & Jordan, M. A. (1993) Cell. Pharma-col., suppl. 1, in press.

31. Reed, E., Sarosy, G., Jamis-Dow, C., Klecher, R., Kohn, E.,Link, C., Christian, M., Davis, P. & Collins, J. (1993) Proc.Am. Assoc. Cancer Res. 34, 395 (abstr.).

32. Speicher, L. A., Barone, L. & Tew, K. D. (1992) Cancer Res.52, 4433-4440.

9556 Cell Biology: Jordan et al.